In this paper a detailed description of collecting Korean audio-video speech data is presented. The main aim of this experiment is to collect some audio-video materials which can be used for later experiments to estimate and model the actions of the visible human articulatory organs such as mouth, lips and jaw. We collect audio-video data from seven directions separately. Twelve markers are used to trace the movements.

The STC(Sinusoidal Transform Coding) is a vocoding technique that uses a sinusoidal speech model to obtain high- quality speech at low data rate. It models and synthesizes the speech signal with fundamental frequency and its harmonic elements in frequency domain. To reduce the data rate, it is necessary to represent the sinusoidal amplitudes and phases with as small number of peaks as possible while maintaining the speech quality. As a basic research to develop a low-rate speech coding algorithm using the sinusoidal model, in this paper, we investigate the speech quality depending on the number of sinusoids. By varying the number of spectral peaks from 5 to 40 speech signals are reconstructed, and then their qualities are evaluated using spectral envelope distortion measure and MOS(Mean Opinion Score). Two approaches are used to obtain the spectral peaks: one is a conventional STFT (Short-Time Fourier Transform), and the other is a multiresolutional analysis method.

This paper concerns effective speaker adaptation methods to solve the over-training problem in speaker verification, which frequently occurs when modeling a speaker with sparse training data. While various speaker adaptations have already been applied to speech recognition, these methods have not yet been formally considered in speaker verification. This paper proposes speaker adaptation methods using a combination of MAP and MLLR adaptations, which are successfully used in speech recognition, and applies to speaker verification. Experimental results show that the speaker verification system using a weighted MAP and MLLR adaptation outperforms that of the conventional speaker models without adaptation by a factor of up to 5 times. From these results, we show that the speaker adaptation method achieves significantly better performance even when only small training data is available for speaker verification.

The present study investigates the clinical applicability of a new device which objectively measures nasal resonating vibration via piezoelectric vibratory sensor from 10 normal volunteers, 10 patients with definite hypernasality and 10 nasal polyposis patients. For the assessment of the hypernasality, the ratio of 'ng' to 'a' as well as that of 'mama' to 'papa' passages were used. For the evaluation of hyponasality, the ratio of nasal vibration post- to pre-induced cul-de-sac resonation was calculated. In the control group, the ratio of ng/a and mama/papa passages was larger than 8, while in the hypernasality group, the ratio was markedly lower. The vibratory signals of 'a' and 'ng' increased markedly in the control group and the hypernasality group after inducing cul-de-sac resonation, while in the hyponasality group, the change was minimal.

Theoretically, alaryngeal speakers have difficulty in accomplishing the production of voiceless consonants. However, the perceptual studies often reveal a clear production of voiceless consonants giving good articulation scores in skilled alaryngeal speakers. The purpose of the present study was to clarify the production of voiceless stops in mode of articulation to normal speakers and skilled alaryngeal speakers. The acoustic characteristics of alaryngeal speech compared to the normal speech were investigated with special reference to the voiceless stop consonants. The surface electromyography from neck is used to monitor pharyngeal activity during speech. The general result is. that esophageal, shunt and neoglottal speakers realize the distinctions between the three types of [p] in a manner parallel to normals, whereas those using an electric voice generator do not.

The present study attempted to determine the effect of olfactory stimulation on word retrieval deficit in Korean aphasics. Eleven nonfluent and 9 fluent aphasic patients served as subjects. The 20 subjects' age ranged from 14 to 65 with the mean of 43.0 (SD: 17.6). A neurologist examined them and diagnosed that they had no visual, auditory, and olfactory impairment. The study consisted of 2 experiments: Experiment I included visual stimulation while experiment II included visual and olfactory stimulation. The subjects were presented with 19 pictures for them to name in the experiment I, whereas they were exposed to the 19 pictures along with the corresponding olfactory stimulus in the experiment II. The 19 items included soy sauce, chilly pepper, Kimchi, chewing gum, cigarette (smoked), soybean paste, lemon, banana, alcohol (wine), apple, fish, vinegar, cuttlefish (dried), milk, rose, sesame oil, melon, coffee, and perfume. The results were as follows: First, olfactory stimulation tended to improve aphasics' word retrieval deficit although the improvement was not statistically significant. Second, the nonfluent aphasics seemed to take more advantage from olfactory stimulation compared to the fluent aphasics. Third, olfactory stimulation (olfactory + visual) did not produce a different naming performance compared to visual stimulation in the pre- and post-test when the pre- and post-test was composed of naming tasks through auditory + visual stimulation. Fourth, the fluent aphasics performed better with unpleasant olfactory stimulus while the nonfluent aphasics performed better with pleasant olfactory stimulus.

Distortion Product Otoacoustic Emissions (DPOAEs) can be measured in the external ear canal two fold: amplitude and latency, but most DPOAE studies deal with amplitude aspects. The purpose of this study was to investigate the latency of the 2f1-f2 DPOAEs in ears with hearing losses and to see if it could be a clinically useful method to distinguish normal from abnormal ears. For this purpose, DPOAE latency were measured as a function of frequency from 1 to 8 kHz in 30 ears with conductive and sensorineural hearing losses (SNHLs). DPOAEs were recorded with Otodynamic Analyzer ILO92. Results showed that the latency decreased as the frequency increased up to 8 kHz. The mean values of DPOAE latency for ears of SNHLs were shorter at all frequencies when they were compared to the mean values of normal ears. The latency in ears of conductive hearing losses was shorter than normal ears at the selective frequencies, as well. The results support the hypothesis that latency values are shorter in pathological ears.

Examination of the vibrations of the vocal fold is very important in patients with voice changes. The newly developed videokymography (VKG) takes images in real time and records irregular vibrations of the vocal fold. However, there are few data on VKG findings. We studied VKG to evaluate the vibratory characteristics of the vocal fold in benign vocal fold lesions. Unique vibratory patterns, blurred demarcation of the mucosal propagation, decreased margin amplitudes, asymmetry in phase or amplitude, and a level difference of the contact site were observed in each benign vocal-fold lesion. We conclude that each type of laryngeal lesion has specific characteristic which are revealed by VKG examination, these parameters can be quantified and used to objectively evaluate VKG findings. Based on these results, VKG can be used as a supplementary diagnostic tool.

Motor aphasia is an affection frequently caused by insult of the left middle cerebral artery and usually accompanied by a large lesion involving the Broca's area and the adjacent motor and premotor areas. Therefore, a patient with motor aphasia commonly shows articulatory disturbances due to failure of the motor programing of speech sound. Objective assessment and treatment of phonologic programing is one of the important aspects of speech therapy in aphasic patients. We analyzed the speech disorders acompanied with motor aphasia in a 45-year-old man using a computerized sound spectrograph, Visi-$Pitch{\circledR}$, and Multi-Dimensional Voice $Program{\circledR}$. We concluded that a computerized speech analysis system is a useful tool to visualize and quantitatively analyse the severity and progression of dysarthria, and the effect of speech therapy.

Acoustic and aerodynamic characteristics of Cheju consonants were examined with the focus on the well-known three-way distinction among stops (i.e., lenis, fortis, aspirated) and the two-way distinction between sand s*. Acoustic parameters examined for the stops included VOT, relative stop burst energy, Fo at the vowel onset, H1-H2, and H1-F2 at the vowel onset. For the fricatives s and s*, acoustic parameters were fricative duration, Fo, centroid of the fricative noise, RMS energy of the frication, H1-H2 and Hl-F2 at the onset of the following vowel. In investigating aerodynamics, intraoral pressure and oral flow were included for the bilabial stops. Results indicate that, although Cheju and Korean are not mutually intelligible, acoustic and aerodynamic properties of Cheju consonants are very similar in every respect to those of the standard Korean. Among other findings there are three crucial points worth recapitulating. First, stops are systematically differentiated by the voice quality of the following vowel. Second, stops are also differentiated by aerodynamic mechanisms. The aspirated and fortis stops are similar in supralaryngeal articulation, but employ a different relation between intraoral pressure and flow. Finally, our study suggests that the fricative s is better categorized as 'lenis' than as 'aspirated' in terms of its phonetic realization.

This chapter presents an overview of Korean intonational structure and proposes a revised version of K -ToBI (Korean TOnes and Break Indices), a prosodic transcription convention for Seoul Korean. In the new version of K-ToBI, a tone tier is separated into two tiers: a phonological tone tier and a phonetic tone tier. A phonological tone tier labels tones marking the prosodic structure of an utterance, and a phonetic tone tier labels individual tones of an AP and an IP conforming to the surface pitch contour. Labelling surface tonal patterns will provide us data to test the underlying tonal patterns and to build phonetic implementation rules.

The aim of this paper is to investigate the acoustic characteristics of the French vowel [e] and Korean [i] and to seek a way of understanding them from a phonological point of view. These two vowels have similar distributional properties, i.e. they alternate with zero in some contexts. Therefore, in both languages, they are not found when immediately followed by a nucleus with phonetic content and in word-final positions. We firstly compare the two vowels by measuring the actual frequencies of the formants, pitch and energy using CSL. We also consider whether the realisation of the two vowels is affected by the speed of speech sounds. In order to show that realisation of the two vowels in both languages is not arbitrary, rather predicted, we will introduce the notion of proper government, proposed and developed by Kaye (1987, 1990) and Charette (1991).

This study was done to investigate the English vowels in relation to the foreign accent. In this study, I also tried to find out the foreign accent of the English diphthong /o/ and front low vowel /${\ae}$/ when Koreans speak English. The reason why I chose these vowels is that these vowels, /o/ and /${\ae}$/, are difficult for Koreans to discern and pronounce. Koreans show a foreign accent in their pronunciation. In order to find out the reason for a Korean foreign accent, experiments were carried out with the help of acoustic instruments. According to the results of the experiment, Koreans showed a foreign accent in their English pronunciation of vowels in relation to their utterance positions. Americans showed a final lengthening effect but Koreans showed a final shortening effect. This means that Koreans showed a foreign accent in the final stressed syllable of a sentence. In addition to this, the duration of two English vowels, /o/ and /${\ae}$/, showed considerably different features between Koreans and Americans. In fact, in the pronunciation of the diphthong /o/, the tongue moves from one position to another. The two articulations of a diphthong can be described as the nucleus plus a glide. However, most Koreans have no idea of this phenomenon and pronounce the diphthongs like two separate monophthongs. This causes a great difference in the lengths of English diphthong /o/ between Koreans and Americans.

The purpose of this study is to revisit the property of the Korean plain stops in intervocalic position. More specifically, focusing on a word-internal, intervocalic position, this study investigates 1) how often speakers pronounce intervocalic. stops as fully voiced, 2) in what amount each speaker voice the plain stops during the stop closure, 3) whether the preceding or the following vowel influences the voicing of target consonants, and 4) the fundamental frequency pattern at the vowel onset after the target consonant shows any consistent pattern, regardless of whether voicing is present during the closure. The results of this study give strong support for the phonetic account of the voicing distinction in Korean. (Jun 1995, 1996).

In this study, I investigate the role of post-consonantal fundamental frequency (F0) as a cue for automatic distinction of types of Korean stops and affricates. Rather than examining data obtained by restricting contexts to a minimum to prevent the interference of irrelevant factors, a relatively natural speaker independent speech corpus is analysed. Automatic and statistical approaches are adopted to annotate data, to minimise speaker variability, and to evaluate the results. In spite of possible loss of information during those automatic analyses, statistics obtained suggest that vowel F0 is a useful cue for distinguishing manners of articulation of Korean non-continuant obstruents having the same place of articulation, especially of lax and aspirated stops and affricates. On the basis of the statistics, automatic classification is attempted over the relevant consonants in a specific context where the micro-prosodic effects appear to be maximised. The results confirm the usefulness of this effect in application for Korean phone recognition.